Supporting O-RAN A1 PoliciesTest & Demo using Non-RealTime RIC functions

Automated test & demo environment

Björn Magnusson (Ericsson)

John Keeney (Ericsson)

The O-RAN A1 Interface

• The A1 interface connects Non-RealTime-RIC logical function in OAM/SMO

layer with the Near-RealTime-RIC logical function in the RAN.

– A1 interface enables policy-based guidance (“A1 Policies”) to be sent to underlying RAN

elements from the management system (“A1-P”).

– “A1-EI” to pass enrichment information from the management platform to the RAN elements

– “A1-ML/AI” to assist with ML Model management in the RAN (Still being defined)

2

High Level Architecture of O-RAN From ORAN-WG1 Architecture Description - v01.00.00 - https://www.o-ran.org/specifications

A1 Policy Functions in ONAP

• A1 Adapter– A1 REST southbound

– RESTCONF Northbound

– Can be included in an any controller based on

ONAP CCSDK

• A1 Policy Management Service– RAN-wide multi-version management of A1 Policy

information

– Query A1 Policy Types in near-RT-RICs

– Status/Create/Query/Update/Delete A1 Policy

Instances in near-RT-RICs

– Optional re-synchronization after inconsistencies or

near-RT-RIC restarts

– Unified REST & DMaaP NBI

– Optionally deploy without A1 Adaptor to connect

direct to near-RT-RICs

Non RealTime RIC Functions in OSC

• A1 Enrichment Information Coordinator

– Registry for A1-EI types, sources, consumers and EI-jobs

– Includes specifics of requested data (e.g. schema), delivery

characteristics, identifier, etc.

– Provides ongoing EI-Job management, LCM & status information

– A1-EI Query API

– Monitors all near-RT-RICs and recovers from inconsistencies

– A1-EI data/flows pass from management layer to RAN directly

• Non-RT RIC Control Panel / GUI

– Query A1 Policy Types supported in RAN

– Model-driven Create/View/Delete/Status of A1 Policy instances

– View registered A1-EI jobs

• (Initial) R-App Register

– Registry for rApps / Services in Non-RT-RIC

– First step towards R-App management

• A1 / near-RT-RIC Simulator

– Used to create multiple stateful A1 providers (simulated near-RT-RICs)

– Supports A1-Policy and A1-Enrichment Information

– Deployment / Configuration API - enables multiple A1 profiles, dynamic metric &

status checks, simulator LCM, A1-P & A1-EI consumer/producer management

• A1 Adapter

– Standalone A1 Controller based on

SDNC

– Also use in combined OSC A1/O1

controller

• A1 Policy Management Service

Non RealTime RIC Functions in OSC

Demo contents

• ONAP – Guilin release– A1 Policy Management Service (A1 Policies)

– SDNC with A1 Adapter (A1-P termination/Mediation)

• O-RAN SC – Cherry Release– A1 Enrichment Coordinator Service (A1 Enrichment data)

– Non-RT-RIC Control Panel (View/Edit A1 Policies & A1 Enrichment Information Jobs)

– A1 / near-RT-RIC Simulator (Terminates A1 Interface)

– R-App Registry (Service Catalog)

– (… and other interface stubs)

• Demo – execute an existing test case – Spin up a new simulated environment - currently docker based

– Add some A1 Policy types, Enable/Define some A1 Policy Types, Create/Deploy some A1 Policies

– Register an A1 Enrichment Information producer and create jobs to send data to the consumer (simulator)

– Register a new R-APP

Test/demo architecture

A1 NonRT RIC Function test env

• A function test environment built around a test engine and support functions

for each component and simulator/interface stub

• Easy and fast to defined new test/demo scenarios

• Based on bash and python scripts

• Multi-platform – MacOS, Ubuntu and Windows (gitbash)

• Requires only docker, docker-compose, python3 and kubernetes

• Tests execute in dockers or kubernetes (but just docker in this demo)

• The test envrionment handles:

– image management

– environment/configuration setup of components

– docker/kubernetes deployment

– all tests - rest calls to/from the components, including checking json data

– test case reporting and detailed logging

– zero touch – start and wait for test completion

Function test/demo setup

Demo time

Summary11

• A1 Policies enables intent-driven and fine-grained updates to RAN using

the O-RAN Non-RealTime RIC function

• Non-RealTime RIC using A1 Policies and A1 Enrichment Information

enables a more intelligent context-aware RAN

• Non-RealTime-RIC, A1 Policies and A1 Enrichment Information opens new

opportunities for RAN optimization and automation.

• Opensource development supporting quality functions for A1, Non-

RealTime RIC and SMO continues in OSC and ONAP

• Ericsson will continue to innovate in Non-RealTime RIC and drive for an

open rApp eco-system.